Learning, Natural and Artificial

The Intelligent Design Series || Part 02

I have never let my schooling interfere with my education.

Mark Twain

Nature has not adapted the young animal to the narrow desk, the crowded curriculum, the silent absorption of complicated facts

John Dewey

Theoretical ideas should always find important applications. This is not an easy doctrine to apply. It contains within itself the problem of keeping knowledge alive, of preventing it from becoming inert, which is the central problem of all education.

Alfred North Whitehead

Here’s an interesting fact: The average English-speaking child learns over ten words per day from the time they start talking all the way through age seventeen (Paul Bloom, How Children Learn the Meaning of Words). I don’t know about you, but I find this amazing. In elementary school, we were required to learn ten vocabulary words a week, and I still remember what an irksome chore that was. It’s amusing to reflect that while that was going on I was learning ten words a day without even noticing.

How do schools manage to take something that is so easy and make it seem so hard? In their paper on situated learning, Seely-Brown, Collins and Duguid note that “learning words from abstract definitions taken out of the context of normal use, the way vocabulary has often been taught, is slow and generally unsuccessful.” Moreover, much of what is learned is “useless in practice,” as evidenced by student-generated examples like, “I was meticulous about falling off the cliff,” or “Mrs. Morrow stimulated the soup.”

Seely-Brown, Collins and Duguid argue that the schools have a similar problem in virtually all subject areas because they “assume a separation between knowing and doing, treating knowledge as an integral, self-sufficient substance, theoretically independent of the situations in which it is learned and used.” This approach is the product of the school’s main teaching paradigm, which is sometimes called the “empty vessel” model: the idea that knowledge can be poured into the minds of students like water into a container. A less formal label for this approach is “telling and testing”.

How bad is the schools’ pedagogical approach? Even ignoring what it does to kill off children’s interest in learning—turning engaged, inquisitive five-year-olds into bored middle-schoolers—school fails on its own terms, in the sense that most of what we “learn” in school is quickly forgotten. To illustrate this, my mentor, Roger Schank used to challenge the audiences in his seminars to take a high school biology test. Over the course of several years, no non-biologist was able to answer more than a couple of questions correctly, even though Schank’s audience typically consisted of professors and graduate students in scientific disciplines at places like Yale, MIT and Stanford.

But it isn’t just esoteric biology that students are forgetting. According to surveys, only 13% of American adults know what a molecule is, only 20% understand that the seasons are not caused by variations in the distance between the earth and the sun, only 33% know that DNA is the carrier of the genetic code, and more than 20% believe that the Sun goes around the Earth. And that’s just in science—equally appalling results have been observed in other domains. For example, over 50% of American 12th graders, asked to identify a world war two ally of the US, picked Germany, Japan, or Italy.
Since everyone studied these things at length in school, we are entitled to ask what went wrong. The answer is that students learn about these topics only in the sense that they are told about them. Where most of these things are concerned, students engage in a minimal amount of meaningful activity—experimentation, analysis, design or realistic problem-solving—that makes the information relevant to them.

School consists largely of an endlessly repeated cycle of telling and testing that works something like this:
First, I tell you a fact. For example, it might be the Stefan-Boltzmann law, which is:


(Where Φ is the energy flux in watts per square meter from a radiating blackbody, σ is Stephan’s constant, which happens to be 5.6704 x 10-8, and T is the object’s temperature in degrees Kelvin.)

Then, I test you on the fact, which looks something like this:

The Stefan-Boltzmann law is:
a) E=MC2
b) F=MA 
c) Φ=σT4
d) All traffic must stop at a red light.

(Did you say c? Very good—you pass!)

And finally, most of the time, you forget the fact, because “knowledge” gained through memorization typically doesn’t last very long. A fact about science—say, “A molecule is the smallest particle of a chemical substance that retains that substance’s chemical composition and properties” —is forgotten right along with the date when Hannibal crossed the Alps, and the name of the author of Jane Eyre. In most cases, this happens soon after the exam is over.

The root of the problem is that learning and memorization are two different things. The human mind is designed to process and retain the lessons of experience, not to remember isolated, de-contextualized facts. This is why a person can talk for hours about what happened to him today while struggling to remember the five facts about the Peloponnesian war he studied last night. It is why a student can learn ten words per day in the course of normal conversation without thinking about it, while struggling to memorize the 10 words per week she must know for the vocabulary test. It is why the standard advice for how to get better at memorizing lists of items is to tell yourself a story that connects those items into some kind of imagined experience. It is, finally, why your ability to remember your own experiences is vastly superior to your ability to remember what you are told.

Because the majority of school lessons give students knowledge that they can recite but not apply in any useful way, most of what students learn is not useful, even when it is retained. Several years ago, Eric Mazur, a physics professor at Harvard, stumbled upon an interesting set of quiz questions created by two science education researchers at Arizona State University. These were simple, intuitively accessible, and entirely number-free questions about the concepts covered in an elementary college physics class. One, for example, noted that the Levi Strauss logo shows two horses attached to a pair of jeans, one per leg, trying to pull them apart, and posed the following question: if Levi had been able to afford only one horse, and had tied the other leg of the pants to a post, would the stress on the pants have been less than, greater than, or equal to the stress with two horses pulling against each other?

Mazur read with astonishment that physics students at a number of big universities had been stumped by many of the questions. He was sure his students would have no such trouble answering them. But, being a good scientist, he put his belief to the test, by adding a couple of the simple, qualitative questions to the next exam he gave his class. He was in for a surprise. The first hint of trouble came when one student raised his hand during the exam and asked: “Professor Mazur, how should I answer these questions? According to what you taught us, or in the way I normally think about these things?”
Mazur’s students did terribly on the qualitative questions, failing to correctly answer them even though most of them correctly answered quantitative versions of what where effectively the same problems elsewhere in the exam. Mazur says, “it became apparent that many students were simply memorizing formulas for solving equations without understanding the underlying physics.” They were, in other words, paying only enough attention to each question to map it to a memorized equation, then calculating the answer by plugging the numbers from the question into that equation—“plugging and chugging,” as generations of science students have called it. The qualitative questions, which in principle should have been much easier to answer, left student without any numerical parameters they could plug into a formula, and this, in turn, exposed their failure to understand the concepts underlying the equations.
(By the way, the answer to the question in the previous paragraph is that the stress on the pants would be the same.)

Mazur realized that, “Conventional lectures simply reinforce students’ feelings that the most important [thing] is memorizing formulas.” His response was radical: he eliminated lectures from his classes. As he put it, lecture and hoping students learn physics from it was a bit like trying to teach students piano by playing for them. He reorganized his classes around discussion and experimentation by the students.
The lesson for the instructional designer in all this seems clear: Don’t teach the way the schools teach! Instead, exploit the mind’s natural learning mechanisms, which are structured to learn from experience. Unfortunately, most instructional design still follows the schools’ tell-and-test approach. This is even true of e-learning: The typical web-based training course is a jazzed-up PowerPoint presentation with quizzes. That is sad, because the computer turns out to be exactly the tool that is needed to get away from the constraints that molded the traditional classroom. When every student has a computer, it becomes possible for each student to engage in a meaningful activity (real or simulated) without massive logistical problems for the instructor. Every computer is potentially a window on the world through which the student can go exploring and learning.

It takes thought and effort to figure out how to engage students’ natural learning mechanisms by designing experiences that teach them the lessons they need to learn. It’s much easier to ask an expert what people need to know about a subject, and then simply tell that to the learners. But it doesn’t work, and the cycle of telling and testing proves nothing except that learns have mastered the skill of cramming for tests. Skills can only be learned through experience, and training can only be made more efficient by understanding and exploiting the way the minds natural learning mechanisms work, not through pointless cycles of memorizing and testing.

Leave a Reply

Your email address will not be published. Required fields are marked *